Method of Ranking in the Function Model

Available online at www.sciencedirect.com

ScienceDirect Procedia CIRP 39 (2016) 22 – 26

TFC 2015 – TRIZ FUTURE 2015

Method of ranking in the function model Nikolai K. Efimov-Soini *, Leonid S. Chechurin Lappeenranta University of Technology, Skinnarilankatu 34, Lappeenranta 53851, Finland * Corresponding author. E-mail address: [email protected]

Abstract At present function analysis is often used for system analysis and concept design development. Function analysis is based on modelling technique and rules of model modification, the most known of which is trimming. Trimming operator suggests system simplifications after each element of it is given a rank. . Thus, the core of the trimming is the evaluation criterion. The article compares two known ranking methods (Gen3 method and method of Miao Li) and suggests a new method of ranking of elements in the function model. Exemplary mechanical system design analysis shows how different ranking approaches influence the trimming procedure. The method can be used for CAD/CAM software at the stage of conceptual design for automatic and semi-automatic system simplification. © 2015 2016 The The Authors. Authors.Published Publishedby byElsevier ElsevierB.V. B.V. This is an open access article under the CC BY-NC-ND license Peer-review under responsibility of scientific committee of Triz Future Conference. (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Scientific committee of Triz Future Conference Keywords: TRIZ, function model, ranking, function analisys,

1. Introduction Methods for systematic conceptual design have always been in the focus of research especially since the whole design became software-frameworked. Obviously, systematic approach means certain formalism of analysis based on modelling and model transformation formalism. One of the most reasonable modelling techniques (neither based on powerful but complicated mathematical nor simple but unformal natural language models) is known to be function modelling [1]. In fact, the usage of function models enables “top-down” and “bottom-top” design style. On the other side, there are CAD system development trends to use TRIZ elements in them, or there are works with the 3D solid body models using TRIZ [2]. Currently there has been some progress in the area automated design tools development. The focus is algorithms and tool for systematic design ideas generation, troubleshooting, design transformation and simplification etc. For example, GoldFire™ software by Invention Machine presents tool to patent design around from [3]. The product supports functional modeling performed by user or even partially automated manner from the text of the patent or

another technical document, then the user performs ranking and the software suggests the elements for trimming. We should notice that big data or more precisely literature based discovery studies attack similar but slightly more general problem. They focus on extraction the concept (e.g. contents, ontology, hierarchy, interactions, subject-object action triples, causeeffect relationships, function model) from the textual data. Another idea to automate the function model design was to use CAD environment, which is standard interface for system data processing in engineering world. The study [2, 4, 5, 6] presents an approach and working prototype of software that automates the function model extraction from 3D SolidWorks CAD assembly, and assist further function ranking and trimming. It is quite possible that the progress in the field will deliver algorithms that are able to design function model (knowledge) of an engineering system from patents, pictures, texts etc. (information). Similar revolution was brought by powerful computers in 90s, when the design of certain types of mathematical models became almost automated due to blending of finite-element approach with graphical system description.

2212-8271 © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of Scientific committee of Triz Future Conference doi:10.1016/j.procir.2016.01.160

Nikolai K. Efimov-Soini and Leonid S. Chechurin / Procedia CIRP 39 (2016) 22 – 26

The goal of the function modelling is to analyze the product we are going to improve. At present, there are great number of methods for assessing the function model such as the solving complexity factor [7], the value engineering [8] etc. All these methods share a common trait to focus on the product model with selected elements only. There are three steps to design the functional model of a system [9]: the component analysis, the interaction analysis, the function analysis. Having designed the function model we can systematically derive models for simplified design by trimming. Let us consider the function ranking and the trimming in detail. 3 (“A, B, C”) or 6 (“A, X, B, C, D, E”) rules are often used for the trimming [10]. It should be noted that these usage is directly related to the rank of the functions. The element with the lowest rank is the first candidate for trimming. The application of the formalized approach simplifies “manual” trimming procedure applications and may serve as the basis for design automation. 2. Description of methods 2.1. Definitions We are going to use the following definitions throughout the paper. x The rank is defined by the ranking factor. The more ranking factor has the higher the rank. x The more useful (or more used) functions (elements) obtain the higher rank, the useless (or unused) functions (elements) obtain the lower rank. x The rank is evaluated by integers from 1 to Ğ, where the function with the highest rank obtains the value 1. So, the higher number has the lower rank. 2.2. Classical method of ranking This method is widely used for systematic inventing [11, 12, 13]. The higher rank belongs to the functions that are closer to the key function in this method. So, we choose the furthest from the target functions as the candidates for trimming. For example, tooth brash bristles are of the highest rank, but the rubber cover on the handle is the lowest rank. Thus, the method may lead to the situation when the highest rank belong to an element that is geometrically close the target but does not perform any special function. For example, a sheet of paper laying on the chair would have the highest rank while adding nothing to main function of the system “to hold”. 2.3. Linear convolution (Method of Miao Li). To evaluate the Function level points (ranking factor) of each component in this method Miao Li [14] introduces the function level score. For example: Useful function (5 point), Harmful function (−5 point), insufficient function (3 point), Excessive function (−3 point). Besides, the importance factor of each function level can be assigned based on expert’s opinion and practical situation.

If one component performs 3 useful functions, 1 harmful function and 2 insufficient functions to other components, and the importance factor of each function level are all equal to 1. Therefore, the component function level points is 16 points (3 ‫ כ‬5 ‫ כ‬1 − 1 ‫כ‬5 ‫ כ‬1 + 2 ‫ כ‬3 ‫ כ‬1). Let us assume the total cost of the system is equal to 100, the cost ratio of this component is 10%. So the component relative cost gets 10. At last, by evaluating each component functionality points (function performance level points over relative cost), the total function rank of the engineering system components can be obtained. The higher score indicates that the component has more functionality. The lower score means that the component has not so much functionality, which gives a higher priority for Trimming. Interestingly, the author prefers using rules A, B, C for the trimming instead of this method [14]. 2.4. New method The above methods have a significant drawback – they are not able to highlight useful elements. Thus we suggest the following approach for ranking. x The closer function is to the target function, the higher is its rank (as the classical method of ranking – 2.2). x The more connections, associating the element with the function, the higher rank each function has. x Duplicate functions obtain the lower rank (for example, 2 nails are holding one board, the function "hold" of each nail has the lower rank). x The farther element is from the key element (geometrically) the lower rate it has. 3. Case study As an example, let us consider the concept, designed to verify the modes of polishing in the TERMOTRONIC firm (St. Petersburg, Russia) [15]. The aim of the development was to check what regimes were the best for polishing of the flowmeter “Piterflow RS” electrodes. This device was not used for industrial electrode polishing, but only to verify the modes of polishing such as the speed handle, the composition of abrasives, the processing time, etc. 3.1. Device description This design of this device was inspired by contact lens polishing system [16, 17]. The device comprises two main systems – a rotation system, and a swing system. We have treated only the swing system by the trimming. The rotation system consists of a spindle (for a hold electrode) and an electric motor rotating a spindle. The swing system design is presented in the figure 1. The main swing system function is to move the mount that sets in motion the pillow with abrasive, polishing the head of the electrode.

23

24

Nikolai K. Efimov-Soini and Leonid S. Chechurin / Procedia CIRP 39 (2016) 22 – 26

Figure 1 - Polishing device 3

3.2. Creation of the function model and the function ranking The function model is shown in the figure 2. Ranking results for three different methods are presented in the tables 1-3. 4

5

6

clip

move

radial bearing

support1

hold

bearing1

support2

hold

bearing2

actuator

hold

clutch

actuator

move

clutch

shaft

hold

clip

shaft

move

clip

support2

hold

plate 2-3

spacer(x3)

hold

actuator

control system

switch

actuator

support3

hold

plate 2-3

support3

hold

spacer(x3)

plate 1-2

hold

control system

support1

hold

plate 1-2

support2

hold

plate 1-2

Table 2. Linear convolution Rank

Element

Ranking factor

1

radial bearing

15

clip

10

support3

10

shaft

5

support1

5

support2

5

plate 1-2

5

control system

5

clutch

0

bearing1

0

bearing2

0

plate 2-3

0

spacer(x3)

0

actuator

0

2

3 Figure 2 – Function model

Table 1. Classical method of ranking Rank 1

Element 1 radial bearing shaft

2

clutch

function distance hold hold

Element2 mount mount shaft

clutch

move

shaft

bearing1

hold

shaft

bearing2

hold

shaft

shaft

hold

radial bearing

clip

hold

radial bearing

4

Nikolai K. Efimov-Soini and Leonid S. Chechurin / Procedia CIRP 39 (2016) 22 – 26 Table 3. Author’s method

4. Conclusions

Rank Element 1

function

Element2

Ranking factor

1

shaft

hold

mount

-3

2

shaft

hold

radial bearing

-2

3

radial bearing

distance

mount

-1A

shaft

hold

clip

-1B

shaft

move

clip

-1B

5

support2

hold

bearing2

0A

6

clutch

move

shaft

0B

actuator

hold

clutch

0C

actuator

move

clutch

0С

8

clip

hold

radial bearing

2A

9

support1

hold

bearing1

2B

10

clutch

hold

shaft

3A

11

support2

hold

plate 2-3

3B

12

control system

switch

actuator

3C

13

spacer(x3)

hold

actuator

3D

bearing1

hold

shaft

4A

4

7

14

bearing2

hold

shaft

4A

15

plate 1-2

hold

control system

4B

16

support2

hold

plate 1-2

5A

15

support3

hold

spacer(x3)

5B

16

support3

hold

plate 2-3

5C

17

support1

hold

plate 1-2

7

For this method we use the principles descripted in 2.4. For example, the function “clutch hold shaft” needs 2 steps to the target function (the initial rank equal 2) ,this one is duplicated from the functions “bearing1 hold shaft” and “bearing2 hold shaft” (2+3), and, as well, the element obtains 2 links (2+32=3). Also the equal factors obtain 4 other functions, but the function “clutch hold shaft” is geometrically closer to the mount (because the elements clutch and the shaft is closer to the mount) than 3 function and is farther than function “support2 hold bearing2”, so this function gets a high rank (B). 3.3. Ranking results In fact the classical method of ranking gives the higher rank to the functions that are the nearest to the target element. For the system we analyze it means the radial bearing (table 1) which is intermediary only. There are the most useless element in the device, in opposite the author’s method “highlights” the most used component. he most useless elements (support1, support2, support3) are defined identically in both methods at the same time. The “useful” element in the linear convolution method (table 2) is same as in the classical method - radial bearing one, but the “useless” element is different. It means that the method is not suitable for the ranking in concepts, but in theory this method “highlights” the most “useless” (but maybe not unused) component, what is very important for the patent design around (tables 1-3).

The paper discusses design complexity reduction algorithms based on function modelling technique. Difference in tow known element ranking methods are highlighted and a new approach is suggested. More precisely, x The classical method of ranking is simple for awareness and understanding, it does not require the specification of carrying out any further calculations, and can be used as a method of rapid assessment, but for the evaluation in the automatic mode, requires the supplementation. x The linear convolution method highlights the most unused or useless item, which is theoretically the easiest to be removed, which is convenient to the patent design around. x The new method as an "upgrade" to the Classical method of ranking, to adapt it to the conduction of the automated trimming. 5. Further development The further development goes in 2 areas: development (and check) the formal mechanism for the trimming after choosing the concept (in the new development) and the development mechanism for automatic and semi-automatic trimming. 6. Acknowledgments The authors would like to acknowledge TEKES, the Finnish Funding Agency for Innovation and its program FiDiPro for the support. References [1] V. Gerasimov, V. Kalish, A. Kuzmin, and S. S. Litvin, Basics of Function-Cost Analysis approach. Guidlines (in Russian). Moscow: Moscow, Inform-FSA, 1991, p. 40. [2] Introducing trimming and function ranking to SolidWorks based on function analysis. Chechurin, Leonid S. Wits, Wessel W. Bakker, Hans M. Vaneker, Tom H.J. Proceedings of Triz Future Conference 2011. 2011. pp. 215-225 [3] https://www.ihs.com/products/design-standards-software-goldfire.html Last visit 2015/06/08[16] [4] Towards multidisciplinary support tools for innovation tasks. W.W. Wits,*, H.M. Bakker, L.S. Chechurin 1st CIRP Global Web Conference: Interdisciplinary Research in Production Engineering. 2012. 9 pages. [5] Integrating TRIZ function modeling in CAD software. Bakker, Hans M. Chechurin, Leonid S. Wits, Wessel W. Proceedings of TRIZfest-2011. 2011. 18 pages. [6] Towards multidisciplinary support tools for innovation tasks. W.W. Wits,*, H.M. Bakker, L.S. Chechurin. 1st CIRP Global Web Conference: Interdisciplinary Research in Production Engineering. 2012. 9 pages. [7] Creating innovative products : using total design : the living legacy of Stuart Pugh / Stuart Pugh. edited by Don Clausing, Ron Andrade. Addison-Wesley, 1996. 544 p. ISBN: 0-201-63485-6 [8] Extended Model for Integrated Value Engineering. Florian G.H. Behncke, Sebastian Maisenbacher, Maik Maurer. Procedia Computer Science Volume 28, 2014, Pages 781–788. [9] TRIZ for Engineers: Enabling Inventive Problem Solving, First Edition. Karen Gadd. 2011 John Wiley & Sons, Ltd. Published 2011 by John Wiley & Sons, Ltd. ISBN:978-0-470-74188-7 р.483 [10] TRIZ-based trimming for process-machine improvements: Slit-valve innovative redesign D. Daniel Sheu, Chun Ting Hou. Computers & Industrial Engineering 66 (2013) p.555–566 [11] Basic GEN3 Innovation Discipline (G3:ID) Training. Pages 168.

25

26

Nikolai K. Efimov-Soini and Leonid S. Chechurin / Procedia CIRP 39 (2016) 22 – 26 [12] Patent “Document semantic analysis/selection with knowledge creativity capability utilizing subject-action-object (SAO) structures” US 6167370 A. https://www.google.ru/patents/US6167370 Last visit 2015/06/04 [13] Patent “Computer based system for imaging and analyzing a process system and indicating values of specific design changes US 6202043 B1” https://www.google.ru/patents/US6202043 Last visit 2015/06/04 [14] A TRIZ-based Trimming method for Patent design around. Miao Li , Xinguo Ming, Lina He, Maokuan Zheng, Zhitao Xu. Computer-Aided Design 62 (2015) pp 20–30.

[15] http://termotronic.ru Last visit 2015/05/20. [16] http://www.sciencechannel.com/tv-shows/how-its-made/videos/how-itsmade-contact-lenses/ Last visit 2015/05/20. [17] Patent “Contact lens polishing system US 3782045 A” https://www.google.ru/patents/US3782045?dq=contact+lenses+polishin g+machine&hl=ru&sa=X&ei=jh9cVb6XJIH7ywPM7YGYDA&ved=0C BwQ6AEwAA Last visit 2015/05/20